timetable performance

[anisometropie]

closes #10377

Performance improvement with a 1000 trains timetable on a short path. Every single train was being re-rendered on every single frame.

improvements:

• virtualizing list of trains to only have elements that are actually displayed in the DOM.
• using useCallback on two props to prevent of re-renders on TimetableTrainCard
  

the 2 useCallbacks on their own prevent any re-render of TimetableTrainCard that are already painted.

adding VList on top of the 2 useCallbacks makes an anonymous wrapper of TimetableTrainCard re-render (internal stuff of the library (z prop)) but it’s still an improvement from using useCallback alone.

Timing

case	timer
base	4:27	https://github.com/user-attachments/assets/23e89906-72c4-43d3-81dc-3336af39c6f6
VList only	1:13	https://github.com/user-attachments/assets/36366f06-c0ce-46c4-aaca-b61f1ea29984
useCallback only	1:25	https://github.com/user-attachments/assets/238d8dc5-a106-4904-a7b3-a0f23671c068
Both	1:00	https://github.com/user-attachments/assets/bbb3e9d6-fa5a-4816-b6d3-432d2a064cd6

a bit of improvement for #8575, but still slow on a click

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.79%. Comparing base (c4c4c58) to head (dd0588c).
Report is 7 commits behind head on dev.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@           Coverage Diff           @@
##              dev   #10371   +/-   ##
=======================================
  Coverage   81.79%   81.79%           
=======================================
  Files        1073     1073           
  Lines      106543   106549    +6     
  Branches      727      727           
=======================================
+ Hits        87147    87154    +7     
+ Misses      19357    19356    -1     
  Partials       39       39           

Flag	Coverage Δ
editoast	74.21% <ø> (+<0.01%)	⬆️
front	89.37% <100.00%> (+<0.01%)	⬆️
gateway	2.18% <ø> (ø)
osrdyne	3.28% <ø> (ø)
railjson_generator	87.50% <ø> (ø)
tests	87.05% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[bloussou]

Nice, will you change the way trains are added in the list (currently 10 after 10) in a following PR ? The backend is ready to accept first train list loading and then the remaining parts and this with parallelized batches.

[RomainValls]

I'm not familiar with the virtualizer solution, but performances and code LGTM

[emersion]

Nice and simple!

Testing with a timetable with 4k trains, the next bottleneck seems to be the conflict list. Real world data seems to contain a lot of these (I get ~15k) and it slows down the whole page significantly.

[emersion]

Other nit for the commit message title: "performance" is uncountable in this context.